Bioinformatics guided synthesis of novel antibiotics from previously inaccessible microbial dark matter

About This Grant

Project Summary Antimicrobial resistance is a growing global threat that is predicted to kill more people than cancer by 2050. The rise in pathogens resistant to our current arsenal of antibiotics necessitates the development of a new generation of therapeutics. Bacterial natural products have historically been the most productive source of mechanistically and structurally novel antibiotics, serving as lead structures for the development of numerous clinically approved therapeutics. It is well established that most of the bacterial biosynthetic diversity on Earth remains locked in microbial dark matter (i.e., unstudied environmental bacteria) that is inaccessible to natural product discovery using existing methods. The inefficiency of previous discovery pipelines results from both the inability to culture most bacteria and the fact that most natural product biosynthetic gene clusters remain silent under laboratory fermentation. In the studies outlined in this proposal, I will couple our new Direct Nanopore Metagenomic Sequencing methods and our Synthetic Bioinformatic Natural Products Discovery methods to overcome both bottlenecks to identify previously inaccessible nature-inspired antibiotics. Direct Nanopore sequencing to acccess previously inaccessble metagenomic biosynthetic gene clusters: One way to overcome issues with culturing bacteria is to sequence DNA directly from environmental samples and identify novel biosynthetic gene clusters from this data. Until now, metagenomic sequencing resulted in assemblies that were too short to fully capture most biosynthetic gene clusters. We have recently developed a robust, direct Nanopore metagenomic sequencing pipeline that readily assembles metagenomic contigs large enough (>100 Kb) to identify complete natural product biosynthetic gene clusters. Accessing antibiotics from metagenomic biosynthetic gene clusters using synthetic bioinformatic natural products: Existing methods for “decoding” biosynthetic gene clusters use biological processes (transcription, translation, biosynthesis). Unfortunately, most biosynthetic gene clusters remain silent under laboratory fermentation. To help address this bottleneck, we have developed a “biology free” biosynthetic gene cluster decoding process where bioinformatics is used to predict the structure encoded by a gene cluster, and then chemical synthesis is used to generate the predicted structure to give compounds we have called synthetic bioinformatic natural products (synBNPs). In our preliminary studies, synBNP antibiotics, like natural products themselves, regularly had novel or rare modes of action and in vivo activity. Coupling these two methods allows us to identify novel compounds from microbial dark matter that has, until now, remained inaccessible using previous natural product discovery methods. This is designed to be an antibiotic discovery proposal. The final output will be a collection of novel, naturally inspired antibiotics that are ready to enter downstream studies to evaluate their potential as clinical candidates.